Karine Truffin

The depletion of fossil fuels and the urgent need to reduce greenhouse gas emissions are driving the development of cleaner internal combustion engines. Beyond electrification, alternative solutions increasingly rely on decarbonized or carbon-free fuels. While promising, these fuels also introduce major challenges, including mixture preparation, increased risks of pre-ignition and flashback, thermo-diffusive instabilities, and increased sensitivity to in-cylinder aerodynamics and thermodynamic conditions. Moreover, advanced engine concepts such as direct injection and hybrid operation require carefully tailored internal flows to control mixing, turbulence intensity, and combustion rates, while avoiding knock, misfire, and pollutant formation. A key scientific challenge is therefore to improve the understanding and control of in-cylinder flows on a cycle-resolved basis and to assess their sensitivity to engine parameters and operating conditions, including highly transient regimes.

Large-eddy simulations (LES) provide a detailed description of flow–combustion interactions and cycle-to-cycle variability (CCV) mechanisms. However, LES remains affected by epistemic uncertainties from various sources and by the high computational cost of multi-cycle simulations. In addition, the specific combustion properties of fuels such as hydrogen and hydrogen–ammonia blends are not yet fully captured by existing combustion models. Although this aspect will not be the focus of this talk, it will be briefly addressed.

Recent advances in data-driven analysis, uncertainty quantification (UQ), and data assimilation offer new perspectives for the design of robust combustion concepts. These approaches enable a systematic treatment of uncertainties related to boundary conditions and engine parameters, improve the predictive capability of numerical simulations, and allow the reproduction of atypical or extreme engine behaviours. This talk discusses emerging opportunities that combine data-driven techniques with LES and experiments to support the development of reliable, high-efficiency engines over a wide operating range. Finally, the proposed approaches extend beyond engine applications and are applicable to a wide range of complex flow problems, including reactive flow simulations in furnaces and gas turbines, wind turbine wake meandering, and atmospheric pollutant dispersion.

Biography

Dr. Truffin is research engineer at IFPEN. Her research focuses on the development of LES and RANS models, and on data-driven methods for the simulation and analysis of turbulent reactive flows in spark-ignition engines and burners. She has also been actively involved in numerous collaborative projects with academic institutions and industrial partners.